| 异构平台下格子Boltzmann方法实现及性能分析 |
| |
| 引用本文： | 张丹丹,徐 莹,徐 磊. 异构平台下格子Boltzmann方法实现及性能分析[J]. 计算机科学, 2012, 39(4): 296-298,303 |
| |
| 作者姓名： | 张丹丹 徐 莹 徐 磊 |
| |
| 作者单位： | 上海超级计算中心 上海201203 |
| |
| 基金项目： | 国家高技术研究发展计划项目，上海市科委科研计划项目课题，上海浦东人才计划项目 |
| |
| 摘 要： | 对CPU+GPU异构平台下的多种并行编程模式进行了研究,并针对格子Boltzmann方法实现了CUDA,MPI+CUDA,MPI+OpenMP+CUDA多级并行算法。结果表明,算法具有较好的加速性能;提出的根据计算量比例参数调节CPU和GPU之间负载均衡的方法,对于在异构平台上实现多级并行处理及资源的有效利用具有一定的参考和应用价值。
|
| 关 键 词： | 异构平台 GPU 格子Boltzmann 并行 |
Performance Analysis and Implementation of Lattice Boltzmann Methods on Heterogeneous Platform |
| |
| ZHANG Dan-dan , XU Ying , XU Lei. Performance Analysis and Implementation of Lattice Boltzmann Methods on Heterogeneous Platform[J]. Computer Science, 2012, 39(4): 296-298,303 |
| |
| Authors: | ZHANG Dan-dan XU Ying XU Lei |
| |
| Affiliation: | (Shanghai Supercomputer Center,Shanghai 201203,China) |
| |
| Abstract: | This work investigated the implementation of lattice Boltzmann method(LBM) with CUDA on the CPU GPU heterogeneous platform with multiple parallel programming models,MPI+CUDA and MPI+OpenMP+CUDA,which shows good performance speedup. A method used to justify the computational load ratio was proposed in this paper to balance the computational time on CPU and GPUs, which provides insightful information about the multi-levelparallclization and efficient usage of different computational resource on the CPU+GPU heterogeneous platform. |
| |
| Keywords: | Heterogeneous platform GPU LBM Parallel |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
点击此处可从《计算机科学》下载全文 |
|
